An end detection algorithm for segmentation of broken and touching characters in handwritten Gurumukhi word

Abstract

Segmentation is one of the important steps of character recognition system. It is an important step because inaccurate segmentation of characters will cause errors in the recognition stage. In optical character Recognition (OCR) system the presence of touching characters and broken characters decreases the accuracy rate of character recognition. Touching of half character or full character with other full character makes the character segmentation very challenging or difficult task. Segmentation of the broken character is quite difficult because vertical profile projection technique assumes the broken parts of the characters as individual characters. In this paper, we have proposed the method of segmentation for touching and broken characters of handwritten Punjabi text that is the Gurumukhi script. The main purpose of this paper is to provide the new segmentation technique based on neighboring pixels for broken characters and increase the accuracy for touching characters.